Share via

What is Azure equivalent for input_audio_noise_reduction ?

Divjot Singh 20 Reputation points
2026-04-28T10:41:37.26+00:00

We have moved from OpenAI deployed Realtime Mini 4o to Azure deployed Realtime 4o Mini. In OpenAI we had input_audio_noise_reduction property. But this property is not available in Azure deployment.

What is the Azure OpenAI equivalent of this property?

Azure Speech in Foundry Tools

2 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. SRILAKSHMI C 18,385 Reputation points Microsoft External Staff Moderator
    2026-04-28T12:27:08.9233333+00:00

    Hello @Divjot Singh

    Thank you for Reaching out to Microsoft Q&A.

    At present, Azure OpenAI Realtime API does not expose a direct equivalent to OpenAI’s input_audio_noise_reduction parameter.

    Why This Difference Exists

    Although Azure OpenAI Realtime and OpenAI’s public Realtime API use the same underlying model family, the available API parameters are not always identical. New capabilities are often introduced in the OpenAI public API first and may be added to Azure OpenAI at a later stage.

    input_audio_noise_reduction is one such parameter. It is currently available in OpenAI’s Realtime API, but it is not yet exposed as a configurable parameter in Azure OpenAI Realtime.

    How Azure OpenAI Handles Audio Today

    Azure OpenAI Realtime automatically performs baseline audio preprocessing as part of the service pipeline. This includes:

    • Background noise suppression
    • Audio normalization
    • Input conditioning for improved transcription and response quality

    This means that a standard level of noise handling is already applied automatically by the service, even though there is no customer-configurable toggle for it.

    Options for Advanced Noise Reduction

    If your application requires more aggressive or customizable noise suppression, the recommended approach is to preprocess the audio before sending it to Azure OpenAI.

    1: Azure Speech SDK

    If you are using the Azure Speech SDK for audio capture or speech processing, it includes built-in audio enhancement capabilities such as:

    • Noise suppression
    • Echo cancellation
    • Automatic gain control

    This is the recommended Microsoft-native solution for client-side audio preprocessing.

    For example, the Speech SDK allows you to configure audio processing options before streaming audio to your application pipeline.

    2: Azure Voice Live SDK

    If you are using the Azure Voice Live SDK for real-time conversational scenarios, it provides an explicit setting for input noise reduction.

    You can configure InputAudioNoiseReduction through VoiceLiveSessionOptions, allowing you to specify the desired level of noise suppression before audio is sent to the model.

    In this case, the functionality is available at the SDK layer rather than directly through the Azure OpenAI Realtime API.

    3: REST API or Custom Clients

    If you are calling Azure OpenAI Realtime directly via REST or a custom WebSocket implementation, there is currently no server-side noise reduction parameter equivalent to input_audio_noise_reduction.

    In this scenario, you should perform audio preprocessing locally before sending audio to Azure OpenAI. Common options include:

    • Azure Speech SDK
    • WebRTC built-in noise suppression
    • Platform-native audio processing APIs
    • Third-party DSP or audio enhancement libraries

    Azure OpenAI Realtime API does not currently provide a direct equivalent to input_audio_noise_reduction.

    Azure OpenAI automatically applies baseline noise suppression and audio conditioning.

    For advanced or customizable noise reduction, audio should be preprocessed on the client side before being sent to Azure OpenAI.

    Please refer this

    Audio processing in Speech SDK: https://docs.microsoft.com/azure/ai-services/speech-service/audio-processing-overview#speech-sdk-integration

    VoiceLiveSessionOptions.InputAudioNoiseReduction: https://learn.microsoft.com/dotnet/api/azure.ai.voicelive.voicelivesessionoptions.inputaudionoisereduction

    Overview of Azure Speech service: https://docs.microsoft.com/azure/cognitive-services/speech-service/overview

    I hope this helps, do let me know if you have further queries.

    Thank you!

    Was this answer helpful?

    1 person found this answer helpful.

  2. AI answer

    2026-04-28T10:41:50.28+00:00

    In Azure OpenAI Realtime, the equivalent capability is still called input_audio_noise_reduction, but it is exposed as part of the Realtime transcription/session configuration rather than as a top‑level OpenAI platform field.

    From the Azure OpenAI in Azure AI Foundry Models REST API (preview) components, input_audio_noise_reduction is defined as:

    • input_audio_noise_reduction: object (can be set to null to turn off)
      • type: near_field or far_field

    This property:

    • Filters audio added to the input audio buffer before it is sent to VAD and the model.
    • Improves VAD and turn detection accuracy and model performance.

    So when moving from OpenAI Realtime Mini 4o to Azure OpenAI Realtime models (for example gpt-4o-realtime-preview or gpt-4o-mini-realtime-preview), configure noise reduction via the input_audio_noise_reduction object in the Realtime session/transcription request body, setting type to near_field (headsets, close mics) or far_field (laptop/conference mics), or null to disable.

    There is no different or renamed Azure‑specific property; the same concept is supported through this input_audio_noise_reduction field in the Azure Realtime API schema.

    References:

    AI-generated content may be incorrect. Read our transparency notes for more information.

    Was this answer helpful?

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.